This paper introduces a flying robot mapping and localization proposal
from an onboard depth camera. The miniature flying robot is part of an ongoing
project related to ambient assisted living and home health. The flying depth camera
is used with a double function; firstly, as a range sensor for mapping from scratch
during navigation, and secondly, as a gray-scale camera for localization. The Harris
corner detection algorithm is implemented as key point detector for the creation
and/or identification of indoor spatial relations. During the localization phase, the
spatial relations created from detected corners in the mapping phase are compared
to the corners identified in the map. The flying robot position is estimated by matching
these spatial relations.